4 files changed, 171 insertions, 35 deletions

diff --git a/net/ipv4/tcp_cong.c b/net/ipv4/tcp_cong.c
index c7cc62c8dc12..e688c687d62d 100644
--- a/net/ipv4/tcp_cong.c
+++ b/net/ipv4/tcp_cong.c
@@ -174,6 +174,34 @@ int tcp_set_congestion_control(struct sock *sk, const char *name)
         return err;
 }
 
+
+/*
+ * Linear increase during slow start
+ */
+void tcp_slow_start(struct tcp_sock *tp)
+{
+        if (sysctl_tcp_abc) {
+                /* RFC3465: Slow Start
+                 * TCP sender SHOULD increase cwnd by the number of
+                 * previously unacknowledged bytes ACKed by each incoming
+                 * acknowledgment, provided the increase is not more than L
+                 */
+                if (tp->bytes_acked < tp->mss_cache)
+                        return;
+
+                /* We MAY increase by 2 if discovered delayed ack */
+                if (sysctl_tcp_abc > 1 && tp->bytes_acked > 2*tp->mss_cache) {
+                        if (tp->snd_cwnd < tp->snd_cwnd_clamp)
+                                tp->snd_cwnd++;
+                }
+        }
+        tp->bytes_acked = 0;
+
+        if (tp->snd_cwnd < tp->snd_cwnd_clamp)
+                tp->snd_cwnd++;
+}
+EXPORT_SYMBOL_GPL(tcp_slow_start);
+
 /*
  * TCP Reno congestion control
  * This is special case used for fallback as well.
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index 981d1203b152..0a461232329f 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -115,8 +115,8 @@ int sysctl_tcp_abc = 1;
 /* Adapt the MSS value used to make delayed ack decision to the 
  * real world.
  */ 
-static inline void tcp_measure_rcv_mss(struct sock *sk,
-                                       const struct sk_buff *skb)
+static void tcp_measure_rcv_mss(struct sock *sk,
+                                const struct sk_buff *skb)
 {
         struct inet_connection_sock *icsk = inet_csk(sk);
         const unsigned int lss = icsk->icsk_ack.last_seg_size; 
@@ -246,8 +246,8 @@ static int __tcp_grow_window(const struct sock *sk, struct tcp_sock *tp,
         return 0;
 }
 
-static inline void tcp_grow_window(struct sock *sk, struct tcp_sock *tp,
-                                   struct sk_buff *skb)
+static void tcp_grow_window(struct sock *sk, struct tcp_sock *tp,
+                            struct sk_buff *skb)
 {
         /* Check #1 */
         if (tp->rcv_ssthresh < tp->window_clamp &&
@@ -341,6 +341,26 @@ static void tcp_clamp_window(struct sock *sk, struct tcp_sock *tp)
                 tp->rcv_ssthresh = min(tp->window_clamp, 2U*tp->advmss);
 }
 
+
+/* Initialize RCV_MSS value.
+ * RCV_MSS is an our guess about MSS used by the peer.
+ * We haven't any direct information about the MSS.
+ * It's better to underestimate the RCV_MSS rather than overestimate.
+ * Overestimations make us ACKing less frequently than needed.
+ * Underestimations are more easy to detect and fix by tcp_measure_rcv_mss().
+ */
+void tcp_initialize_rcv_mss(struct sock *sk)
+{
+        struct tcp_sock *tp = tcp_sk(sk);
+        unsigned int hint = min_t(unsigned int, tp->advmss, tp->mss_cache);
+
+        hint = min(hint, tp->rcv_wnd/2);
+        hint = min(hint, TCP_MIN_RCVMSS);
+        hint = max(hint, TCP_MIN_MSS);
+
+        inet_csk(sk)->icsk_ack.rcv_mss = hint;
+}
+
 /* Receiver "autotuning" code.
  *
  * The algorithm for RTT estimation w/o timestamps is based on
@@ -735,6 +755,27 @@ __u32 tcp_init_cwnd(struct tcp_sock *tp, struct dst_entry *dst)
         return min_t(__u32, cwnd, tp->snd_cwnd_clamp);
 }
 
+/* Set slow start threshold and cwnd not falling to slow start */
+void tcp_enter_cwr(struct sock *sk)
+{
+        struct tcp_sock *tp = tcp_sk(sk);
+
+        tp->prior_ssthresh = 0;
+        tp->bytes_acked = 0;
+        if (inet_csk(sk)->icsk_ca_state < TCP_CA_CWR) {
+                tp->undo_marker = 0;
+                tp->snd_ssthresh = inet_csk(sk)->icsk_ca_ops->ssthresh(sk);
+                tp->snd_cwnd = min(tp->snd_cwnd,
+                                   tcp_packets_in_flight(tp) + 1U);
+                tp->snd_cwnd_cnt = 0;
+                tp->high_seq = tp->snd_nxt;
+                tp->snd_cwnd_stamp = tcp_time_stamp;
+                TCP_ECN_queue_cwr(tp);
+
+                tcp_set_ca_state(sk, TCP_CA_CWR);
+        }
+}
+
 /* Initialize metrics on socket. */
 
 static void tcp_init_metrics(struct sock *sk)
@@ -2070,8 +2111,8 @@ static inline void tcp_ack_update_rtt(struct sock *sk, const int flag,
                 tcp_ack_no_tstamp(sk, seq_rtt, flag);
 }
 
-static inline void tcp_cong_avoid(struct sock *sk, u32 ack, u32 rtt,
-                                  u32 in_flight, int good)
+static void tcp_cong_avoid(struct sock *sk, u32 ack, u32 rtt,
+                           u32 in_flight, int good)
 {
         const struct inet_connection_sock *icsk = inet_csk(sk);
         icsk->icsk_ca_ops->cong_avoid(sk, ack, rtt, in_flight, good);
@@ -2082,7 +2123,7 @@ static inline void tcp_cong_avoid(struct sock *sk, u32 ack, u32 rtt,
  * RFC2988 recommends to restart timer to now+rto.
  */
 
-static inline void tcp_ack_packets_out(struct sock *sk, struct tcp_sock *tp)
+static void tcp_ack_packets_out(struct sock *sk, struct tcp_sock *tp)
 {
         if (!tp->packets_out) {
                 inet_csk_clear_xmit_timer(sk, ICSK_TIME_RETRANS);
@@ -2147,7 +2188,7 @@ static int tcp_tso_acked(struct sock *sk, struct sk_buff *skb,
         return acked;
 }
 
-static inline u32 tcp_usrtt(const struct sk_buff *skb)
+static u32 tcp_usrtt(const struct sk_buff *skb)
 {
         struct timeval tv, now;
 
@@ -2583,8 +2624,8 @@ void tcp_parse_options(struct sk_buff *skb, struct tcp_options_received *opt_rx,
 /* Fast parse options. This hopes to only see timestamps.
  * If it is wrong it falls back on tcp_parse_options().
  */
-static inline int tcp_fast_parse_options(struct sk_buff *skb, struct tcphdr *th,
-                                         struct tcp_sock *tp)
+static int tcp_fast_parse_options(struct sk_buff *skb, struct tcphdr *th,
+                                  struct tcp_sock *tp)
 {
         if (th->doff == sizeof(struct tcphdr)>>2) {
                 tp->rx_opt.saw_tstamp = 0;
@@ -2804,8 +2845,7 @@ static void tcp_fin(struct sk_buff *skb, struct sock *sk, struct tcphdr *th)
         }
 }
 
-static __inline__ int
-tcp_sack_extend(struct tcp_sack_block *sp, u32 seq, u32 end_seq)
+static inline int tcp_sack_extend(struct tcp_sack_block *sp, u32 seq, u32 end_seq)
 {
         if (!after(seq, sp->end_seq) && !after(sp->start_seq, end_seq)) {
                 if (before(seq, sp->start_seq))
@@ -2817,7 +2857,7 @@ tcp_sack_extend(struct tcp_sack_block *sp, u32 seq, u32 end_seq)
         return 0;
 }
 
-static inline void tcp_dsack_set(struct tcp_sock *tp, u32 seq, u32 end_seq)
+static void tcp_dsack_set(struct tcp_sock *tp, u32 seq, u32 end_seq)
 {
         if (tp->rx_opt.sack_ok && sysctl_tcp_dsack) {
                 if (before(seq, tp->rcv_nxt))
@@ -2832,7 +2872,7 @@ static inline void tcp_dsack_set(struct tcp_sock *tp, u32 seq, u32 end_seq)
         }
 }
 
-static inline void tcp_dsack_extend(struct tcp_sock *tp, u32 seq, u32 end_seq)
+static void tcp_dsack_extend(struct tcp_sock *tp, u32 seq, u32 end_seq)
 {
         if (!tp->rx_opt.dsack)
                 tcp_dsack_set(tp, seq, end_seq);
@@ -2890,7 +2930,7 @@ static void tcp_sack_maybe_coalesce(struct tcp_sock *tp)
         }
 }
 
-static __inline__ void tcp_sack_swap(struct tcp_sack_block *sack1, struct tcp_sack_block *sack2)
+static inline void tcp_sack_swap(struct tcp_sack_block *sack1, struct tcp_sack_block *sack2)
 {
         __u32 tmp;
 
@@ -3455,7 +3495,7 @@ void tcp_cwnd_application_limited(struct sock *sk)
         tp->snd_cwnd_stamp = tcp_time_stamp;
 }
 
-static inline int tcp_should_expand_sndbuf(struct sock *sk, struct tcp_sock *tp)
+static int tcp_should_expand_sndbuf(struct sock *sk, struct tcp_sock *tp)
 {
         /* If the user specified a specific send buffer setting, do
          * not modify it.
@@ -3502,7 +3542,7 @@ static void tcp_new_space(struct sock *sk)
         sk->sk_write_space(sk);
 }
 
-static inline void tcp_check_space(struct sock *sk)
+static void tcp_check_space(struct sock *sk)
 {
         if (sock_flag(sk, SOCK_QUEUE_SHRUNK)) {
                 sock_reset_flag(sk, SOCK_QUEUE_SHRUNK);
@@ -3512,7 +3552,7 @@ static inline void tcp_check_space(struct sock *sk)
         }
 }
 
-static __inline__ void tcp_data_snd_check(struct sock *sk, struct tcp_sock *tp)
+static inline void tcp_data_snd_check(struct sock *sk, struct tcp_sock *tp)
 {
         tcp_push_pending_frames(sk, tp);
         tcp_check_space(sk);
@@ -3544,7 +3584,7 @@ static void __tcp_ack_snd_check(struct sock *sk, int ofo_possible)
         }
 }
 
-static __inline__ void tcp_ack_snd_check(struct sock *sk)
+static inline void tcp_ack_snd_check(struct sock *sk)
 {
         if (!inet_csk_ack_scheduled(sk)) {
                 /* We sent a data segment already. */
@@ -3692,8 +3732,7 @@ static int __tcp_checksum_complete_user(struct sock *sk, struct sk_buff *skb)
         return result;
 }
 
-static __inline__ int
-tcp_checksum_complete_user(struct sock *sk, struct sk_buff *skb)
+static inline int tcp_checksum_complete_user(struct sock *sk, struct sk_buff *skb)
 {
         return skb->ip_summed != CHECKSUM_UNNECESSARY &&
                 __tcp_checksum_complete_user(sk, skb);
@@ -4474,3 +4513,4 @@ EXPORT_SYMBOL(sysctl_tcp_abc);
 EXPORT_SYMBOL(tcp_parse_options);
 EXPORT_SYMBOL(tcp_rcv_established);
 EXPORT_SYMBOL(tcp_rcv_state_process);
+EXPORT_SYMBOL(tcp_initialize_rcv_mss);
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index 9b62d80bb20f..5c70493dff02 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -270,8 +270,7 @@ failure:
 /*
  * This routine does path mtu discovery as defined in RFC1191.
  */
-static inline void do_pmtu_discovery(struct sock *sk, struct iphdr *iph,
-                                     u32 mtu)
+static void do_pmtu_discovery(struct sock *sk, struct iphdr *iph, u32 mtu)
 {
         struct dst_entry *dst;
         struct inet_sock *inet = inet_sk(sk);
@@ -662,7 +661,7 @@ static void tcp_v4_reqsk_destructor(struct request_sock *req)
         kfree(inet_rsk(req)->opt);
 }
 
-static inline void syn_flood_warning(struct sk_buff *skb)
+static void syn_flood_warning(struct sk_buff *skb)
 {
         static unsigned long warntime;
 
@@ -677,8 +676,8 @@ static inline void syn_flood_warning(struct sk_buff *skb)
 /*
  * Save and compile IPv4 options into the request_sock if needed.
  */
-static inline struct ip_options *tcp_v4_save_options(struct sock *sk,
-                                                     struct sk_buff *skb)
+static struct ip_options *tcp_v4_save_options(struct sock *sk,
+                                              struct sk_buff *skb)
 {
         struct ip_options *opt = &(IPCB(skb)->opt);
         struct ip_options *dopt = NULL;
diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
index 3a0a914de917..a7623ead39a8 100644
--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -51,8 +51,8 @@ int sysctl_tcp_retrans_collapse = 1;
  */
 int sysctl_tcp_tso_win_divisor = 3;
 
-static inline void update_send_head(struct sock *sk, struct tcp_sock *tp,
-                                    struct sk_buff *skb)
+static void update_send_head(struct sock *sk, struct tcp_sock *tp,
+                             struct sk_buff *skb)
 {
         sk->sk_send_head = skb->next;
         if (sk->sk_send_head == (struct sk_buff *)&sk->sk_write_queue)
@@ -124,8 +124,8 @@ static void tcp_cwnd_restart(struct sock *sk, struct dst_entry *dst)
         tp->snd_cwnd_used = 0;
 }
 
-static inline void tcp_event_data_sent(struct tcp_sock *tp,
-                                       struct sk_buff *skb, struct sock *sk)
+static void tcp_event_data_sent(struct tcp_sock *tp,
+                                struct sk_buff *skb, struct sock *sk)
 {
         struct inet_connection_sock *icsk = inet_csk(sk);
         const u32 now = tcp_time_stamp;
@@ -142,7 +142,7 @@ static inline void tcp_event_data_sent(struct tcp_sock *tp,
                 icsk->icsk_ack.pingpong = 1;
 }
 
-static __inline__ void tcp_event_ack_sent(struct sock *sk, unsigned int pkts)
+static inline void tcp_event_ack_sent(struct sock *sk, unsigned int pkts)
 {
         tcp_dec_quickack_mode(sk, pkts);
         inet_csk_clear_xmit_timer(sk, ICSK_TIME_DACK);
@@ -212,7 +212,7 @@ void tcp_select_initial_window(int __space, __u32 mss,
  * value can be stuffed directly into th->window for an outgoing
  * frame.
  */
-static __inline__ u16 tcp_select_window(struct sock *sk)
+static u16 tcp_select_window(struct sock *sk)
 {
         struct tcp_sock *tp = tcp_sk(sk);
         u32 cur_win = tcp_receive_window(tp);
@@ -250,6 +250,75 @@ static __inline__ u16 tcp_select_window(struct sock *sk)
         return new_win;
 }
 
+static void tcp_build_and_update_options(__u32 *ptr, struct tcp_sock *tp,
+                                         __u32 tstamp)
+{
+        if (tp->rx_opt.tstamp_ok) {
+                *ptr++ = __constant_htonl((TCPOPT_NOP << 24) |
+                                          (TCPOPT_NOP << 16) |
+                                          (TCPOPT_TIMESTAMP << 8) |
+                                          TCPOLEN_TIMESTAMP);
+                *ptr++ = htonl(tstamp);
+                *ptr++ = htonl(tp->rx_opt.ts_recent);
+        }
+        if (tp->rx_opt.eff_sacks) {
+                struct tcp_sack_block *sp = tp->rx_opt.dsack ? tp->duplicate_sack : tp->selective_acks;
+                int this_sack;
+
+                *ptr++ = htonl((TCPOPT_NOP  << 24) |
+                               (TCPOPT_NOP  << 16) |
+                               (TCPOPT_SACK <<  8) |
+                               (TCPOLEN_SACK_BASE + (tp->rx_opt.eff_sacks *
+                                                     TCPOLEN_SACK_PERBLOCK)));
+                for(this_sack = 0; this_sack < tp->rx_opt.eff_sacks; this_sack++) {
+                        *ptr++ = htonl(sp[this_sack].start_seq);
+                        *ptr++ = htonl(sp[this_sack].end_seq);
+                }
+                if (tp->rx_opt.dsack) {
+                        tp->rx_opt.dsack = 0;
+                        tp->rx_opt.eff_sacks--;
+                }
+        }
+}
+
+/* Construct a tcp options header for a SYN or SYN_ACK packet.
+ * If this is every changed make sure to change the definition of
+ * MAX_SYN_SIZE to match the new maximum number of options that you
+ * can generate.
+ */
+static void tcp_syn_build_options(__u32 *ptr, int mss, int ts, int sack,
+                                  int offer_wscale, int wscale, __u32 tstamp,
+                                  __u32 ts_recent)
+{
+        /* We always get an MSS option.
+         * The option bytes which will be seen in normal data
+         * packets should timestamps be used, must be in the MSS
+         * advertised.  But we subtract them from tp->mss_cache so
+         * that calculations in tcp_sendmsg are simpler etc.
+         * So account for this fact here if necessary.  If we
+         * don't do this correctly, as a receiver we won't
+         * recognize data packets as being full sized when we
+         * should, and thus we won't abide by the delayed ACK
+         * rules correctly.
+         * SACKs don't matter, we never delay an ACK when we
+         * have any of those going out.
+         */
+        *ptr++ = htonl((TCPOPT_MSS << 24) | (TCPOLEN_MSS << 16) | mss);
+        if (ts) {
+                if(sack)
+                        *ptr++ = __constant_htonl((TCPOPT_SACK_PERM << 24) | (TCPOLEN_SACK_PERM << 16) |
+                                                  (TCPOPT_TIMESTAMP << 8) | TCPOLEN_TIMESTAMP);
+                else
+                        *ptr++ = __constant_htonl((TCPOPT_NOP << 24) | (TCPOPT_NOP << 16) |
+                                                  (TCPOPT_TIMESTAMP << 8) | TCPOLEN_TIMESTAMP);
+                *ptr++ = htonl(tstamp);         /* TSVAL */
+                *ptr++ = htonl(ts_recent);      /* TSECR */
+        } else if(sack)
+                *ptr++ = __constant_htonl((TCPOPT_NOP << 24) | (TCPOPT_NOP << 16) |
+                                          (TCPOPT_SACK_PERM << 8) | TCPOLEN_SACK_PERM);
+        if (offer_wscale)
+                *ptr++ = htonl((TCPOPT_NOP << 24) | (TCPOPT_WINDOW << 16) | (TCPOLEN_WINDOW << 8) | (wscale));
+}
 
 /* This routine actually transmits TCP packets queued in by
  * tcp_do_sendmsg().  This is used by both the initial
@@ -724,7 +793,7 @@ unsigned int tcp_current_mss(struct sock *sk, int large_allowed)
 
 /* Congestion window validation. (RFC2861) */
 
-static inline void tcp_cwnd_validate(struct sock *sk, struct tcp_sock *tp)
+static void tcp_cwnd_validate(struct sock *sk, struct tcp_sock *tp)
 {
         __u32 packets_out = tp->packets_out;
 
@@ -773,7 +842,7 @@ static inline unsigned int tcp_cwnd_test(struct tcp_sock *tp, struct sk_buff *sk
 /* This must be invoked the first time we consider transmitting
  * SKB onto the wire.
  */
-static inline int tcp_init_tso_segs(struct sock *sk, struct sk_buff *skb, unsigned int mss_now)
+static int tcp_init_tso_segs(struct sock *sk, struct sk_buff *skb, unsigned int mss_now)
 {
         int tso_segs = tcp_skb_pcount(skb);
 
@@ -1794,7 +1863,7 @@ struct sk_buff * tcp_make_synack(struct sock *sk, struct dst_entry *dst,
 /* 
  * Do all connect socket setups that can be done AF independent.
  */ 
-static inline void tcp_connect_init(struct sock *sk)
+static void tcp_connect_init(struct sock *sk)
 {
         struct dst_entry *dst = __sk_dst_get(sk);
         struct tcp_sock *tp = tcp_sk(sk);